Ambient panel temperature effect on production

[bkp_duke]

Just noticed something very interesting on today's solar production that I thought I would share.

As most here are aware, solar panels are most efficient when cool. There is a temperature coefficient rating for panels that determines how much production they lose with each degree C increase in panel (not ambient) temperature.

For example, I have SunPower X21-335-BLK panels, and these panels lose 0.29% of production for every 1 degree C above 25C (https://us.sunpower.com/sites/defau...x21-335-blk-x21-345-datasheet-504828-revg.pdf)


Like many residents in SoCal, we experience heavy marine layer effects in the mornings. This morning it was relatively cold (low 60s), but the layer broke very quickly. This resulted in a period of time where the panels were cool, but the sun was bright. As the panels heated up, their production dropped, even though the azimuth of the sun was increasing, and production would be expected to increase (see red oval in image below).

Just thought I would share this, in case anyone else sees something similar and wonders what is going on.

 

[Alonzo]

Pardon my ignorance, but how does something black and mounted on a roof on a 95F day not get extremely hot? Isn't that the whole principle of those roof mounted water heating systems we saw back in the 70's?

 

[bkp_duke]

Pardon my ignorance, but how does something black and mounted on a roof on a 95F day not get extremely hot? Isn't that the whole principle of those roof mounted water heating systems we saw back in the 70's?

Of course it gets hot, but people don't usually understand that the "rated output" of solar panels is when their temperature is at 25 C. That's the standard NREL testing setup, and the temperature at which the output and efficiency are determined.

Because, as you point out, the standard operating temperature, especially in the summer months, is much much higher, people will see MARKED reductions in production.

Basically, it explains why a 14 kw DC system would be producing only 12.5 kw DC (before inverter losses with the conversion to AC).



And I posted the graph because we RARELY see this effect in measurable data, but I saw it today because of rapid cloud clearing.

 

[aesculus]

Pardon my ignorance, but how does something black and mounted on a roof on a 95F day not get extremely hot? Isn't that the whole principle of those roof mounted water heating systems we saw back in the 70's?

Absolutely.

But I have a free standing solar unit that has airflow all around it and the panels get hot and degrade just as @bkp_duke talks about. I can lose 10% or more of my production to heat. Panels are rated at 75F so when you operated them at 95-110F, they lose a lot of the potential.

There is a whole science being developed out there for both active and passive cooling of solar panels, especially for adoption in high temperature climates.

 

[aesculus]

And I posted the graph because we RARELY see this effect in measurable data, but I saw it today because of rapid cloud clearing.

And that graph: What was your data source for that? Did you download the data from the Tesla App?

 

[bkp_duke]

And that graph: What was your data source for that? Did you download the data from the Tesla App?

pvoutput.org. I used the instructions in this blog post (Monitoring Tesla’s Powerwall2 on PVOutput.org – MikesGear) to setup a service running on my home server that polls the Powerwall gateway every 5 min, and uploads the data to pvoutput.org.

 

[charlesj]

How would one measure panel temperatures? Even with Enphase inverters that can show temperature, I would think their data is for their inverter temperature, not the panel temp. There are only two cables coming from the panel for the DC volt and current. How does that correlate to temperature?
In a lab, one can actually attach a temp sensor to a panel and see what happens.

 

[bkp_duke]

How would one measure panel temperatures? Even with Enphase inverters that can show temperature, I would think their data is for their inverter temperature, not the panel temp. There are only two cables coming from the panel for the DC volt and current. How does that correlate to temperature?
In a lab, one can actually attach a temp sensor to a panel and see what happens.

FLIR

 

[charlesj]

FLIR

Thanks. Will rush out and get the $700 phone camera.

 

[bkp_duke]

Reproducible again today. Cold AM clouds, quick clearing.

 

[wwu123]

Reproducible again today. Cold AM clouds, quick clearing.

Cool to see. When you mentally superimpose the normal curve from a cloudless day, one can see first the dip (due to clouds), the overproduction spike, and then the quick reversion back (and down) to the normal curve.

The other two times you briefly see this type of spike are 1) when washing off the panels on a sunny day (not advised), and 2) cloud edge effects.

Aside from short-lived temperature effects, I see as much as 8% difference in power on adjacent panels, largely due to temperature, based on where in the array they are situated. The ones at the edge than in the middle, the edges closer to an eave or a hip even cooler.

 

[bkp_duke]

Cool to see. When you mentally superimpose the normal curve from a cloudless day, one can see first the dip (due to clouds), the overproduction spike, and then the quick reversion back (and down) to the normal curve.

The other two times you briefly see this type of spike are 1) when washing off the panels on a sunny day (not advised), and 2) cloud edge effects.

Aside from short-lived temperature effects, I see as much as 8% difference in power on adjacent panels, largely due to temperature, based on where in the array they are situated. The ones at the edge than in the middle, the edges closer to an eave or a hip even cooler.

Agreed. I haven't seen "cloud edge" effects, but panel temperature is definitely noticable.

On a 12.06 kw (DC) system, the highest production I have ever seen was 11.1 kw, and that was on a day in April. In the clear summer months, we never get above about 9.5 kw due to panel heat.


And for sure don't wash those panels in the hot sun! Tempered glass or not, you are risking a cracked panel from the rapid cooling effect.

 

[JayClark]

Very interesting information. I've informally seen some similar effects here in Phoenix, where on a hot day (110+), a cloud passes over for a minute or so (just watching the real-time local gateway production animation) the production will drop as expected, but when the cloud clears the production jumps 250-500 watts higher than what it was previously producing for a few minutes until the panel reheats, and then drops off again, and repeats with the next cloud that comes. (this is on a 4.1kW PV system) . My inverter doesn't cap at max PV production just because of the way the sizing worked out, so these little production spikes jumped out to me because they were up to 400 watts higher than what I had been seeing as the normal charted production high on the production curve on what are usually totally clear summer days for my system. This was a clear-blue sky day, but with scattered thick clouds passing by periodically.

So Similar curve to what you're seeing, but on smaller time increments as clouds come and go.

 

[wjgjr]

Very interesting information. I've informally seen some similar effects here in Phoenix, where on a hot day (110+), a cloud passes over for a minute or so (just watching the real-time local gateway production animation) the production will drop as expected, but when the cloud clears the production jumps 250-500 watts higher than what it was previously producing for a few minutes until the panel reheats, and then drops off again, and repeats with the next cloud that comes. (this is on a 4.1kW PV system) . My inverter doesn't cap at max PV production just because of the way the sizing worked out, so these little production spikes jumped out to me because they were up to 400 watts higher than what I had been seeing as the normal charted production high on the production curve on what are usually totally clear summer days for my system. This was a clear-blue sky day, but with scattered thick clouds passing by periodically.

So Similar curve to what you're seeing, but on smaller time increments as clouds come and go.

Just to be clear, the cloud edge effect is attributed to sunlight reflecting off the edge of the clouds, thus temporarily increasing the solar energy focused on your array. It is not due (at least primarily) to panel cooling/heating, particularly since this phenomenon is often noted with lighter, scattered clouds where the panels would not have much time to cool.

 

[BrettS]

Agreed. I haven't seen "cloud edge" effects, but panel temperature is definitely noticable.

The below graph shows my most noticeable cloud edge effect. It was a pretty cloudless day, but right in the middle you can see the lowered production when some clouds formed and then a spike as they blew away and the cloud edge effect happened. If you look closely you can even see a little increase in production right before the dip, which I think was the cloud edge effect as the sun started to get covered by the clouds.

 

[Alonzo]

And for sure don't wash those panels in the hot sun! Tempered glass or not, you are risking a cracked panel from the rapid cooling effect.

I'm surprised by this. I live in SE Texas, where all through June, July, and August it is in mid to high 90's. We get sudden showers daily, sometimes while it is still sun shining. Are the panels so fragile that this might break them?

 

[bkp_duke]

I'm surprised by this. I live in SE Texas, where all through June, July, and August it is in mid to high 90's. We get sudden showers daily, sometimes while it is still sun shining. Are the panels so fragile that this might break them?

From rain, probably not. But if you ask all the manufacturers, they will explicitly tell you not to clean your panels when they are hot, there there is a small but non-zero chance of cracking from the rapid temperature change.

 

[JayClark]

Just to be clear, the cloud edge effect is attributed to sunlight reflecting off the edge of the clouds, thus temporarily increasing the solar energy focused on your array. It is not due (at least primarily) to panel cooling/heating, particularly since this phenomenon is often noted with lighter, scattered clouds where the panels would not have much time to cool.

Ok, gotcha. Yeah, I think the effect I’m talking about is not that at all. What I was observing was more related to significant panel cooling in our extreme desert heat conditions during the peak solar part of the day, and getting a spike from that rather than from the cloud edge effect.

 

[JayClark]

From rain, probably not. But if you ask all the manufacturers, they will explicitly tell you not to clean your panels when they are hot, there there is a small but non-zero chance of cracking from the rapid temperature change.

I’ve never seen it with panels, but I know here in the desert I have seen a car windshield shatter when someone was watering plants and turned it onto a car window. Window popped and complete shattered - So I would never spray water onto a piece of hot piece of glass.

I’m guessing in most cases with rain, it would start with some light sprinkles, and maybe some breeze or wind, and then build up slowly from there where it’s rarely A problem, that might might be enough to avoid the panels cracking in most cases when it comes to rain.

But suddenly turning a hose full of relatively cool water on a very hot panel, I could see it cracking, and clearly the manufactures have seen it.